EVALUATION OF DESIGN FORCE COEFFICIENTS ON TALL BUILDING CONSIDERING NON-EXCEEDANCE PROBABILITY

Abstract

　In current Recommendations for Loads on Buildings (Architectural Institute of Japan, 2015) (abbreviated as AIJ-RLB (2015)), wind loads are determined based on the concept of an equivalent static wind load, and structural frames are assumed to behave elastically in strong wind. The along-wind load is generally composed of a mean component caused by the mean wind speed, a quasi-static component caused by relatively low frequency fluctuation and the 1st mode resonant component caused by fluctuation in the vicinity of the natural frequency. In AIJ-RLB (2015), the procedure that can estimate the equivalent static wind load producing the maximum structural responses using the gust effect factor was prescribed. For along-wind load, various force coefficients (in AIJ-RLB (2015), wind force coefficient at the top of the building, a factor relevant to overturning moment in the along-wind direction and a factor relevant to rms overturning moment in the along-wind direction), which were used to calculate the mean load and gust effect factor, were modeled as a functions of building shapes and approaching flow characteristics. The across-wind load can be obtained by multiplying fluctuating overturning moment coefficients by the peak factor and the fluctuating overturning moment coefficients were modeled as a function of side ratio only.

　The wind force coefficients in AIJ-RLB (2015) are prescribed based on the wind tunnel tests as an ensemble average of ten and/or twenty 10-minute samples. The small number of 10-minute samples makes it difficult, almost impossible, to identify the exact probability distribution and exceedance probability of the prescribed force coefficients.

　In the present study, more than 10,000 10-munute samples on tall buildings were measured by wind tunnel tests, and the probability distributions of local force and overturning moment coefficients, which were used for the evaluation of the horizontal wind loads on structural frame, were identified first and the design local force and overturning moment coefficients corresponding to the specific non-exceedance probabilities of wind load were calculated based on the probabilistic theory.

　The following concluding remarks can be made from the calculation results.

1) It was possible to identify the probability distributions and non-exceedance probabilities of various wind force coefficients prescribed in AIJ-RLB (2015).

2) The coefficient of variance of wind force coefficients depends on the force components, generally showing larger coefficient of variance assumed in AIJ-RLB (2015).

3) The probability distributions of all the mean and fluctuating force coefficients were found to be normal distribution. And the ratio of the design force coefficient to the arithmetic mean of 10-minutes samples was less than 2%.

4) Considering the arithmetic mean of 10-minute samples corresponds to the value in the AIJ-RLB (2015), the non-exceedance probability of the force coefficients prescribed in AIJ-RLB (2015) could be interpreted to be similar to 99.9%.